Semantic clustering tool for finding patterns in AI-generated code review feedback.

• Embeds text using sentence-transformers (all-MiniLM-L6-v2)
• Computes pairwise cosine similarity
• Clusters using three methods: graph-based, agglomerative, HDBSCAN
• Visualizes with t-SNE projections
• Includes threshold sweep for parameter tuning

Built to identify duplicate/similar issues in large sets of AI-generated code review comments. Reduces noise, surfaces patterns.

python main.py "data/tasks.*.md" --cluster-threshold 0.75 --sweep --dump

Python, NumPy, scikit-learn, HDBSCAN, NetworkX, sentence-transformers, matplotlib